---
base_model: teknium/OpenHermes-2.5-Mistral-7B
tags:
- mistral
- instruct
- finetune
- chatml
- gpt4
- synthetic data
- distillation
- dpo
- rlhf
- exl2
license: apache-2.0
language:
- en
datasets:
- mlabonne/chatml_dpo_pairs
---

EXL2 quantisation of NeuralHermes-2.5-Mistral-7B, for use with ExLLamaV2.

[Original model](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B) by @mlabonne.

**Model size:** 4.6GB (3x reduction), 5 bits-per-weight average, 6bpw on head.

**Calibration Data:** Wikitext [(parquet)](https://huggingface.co/datasets/wikitext/blob/refs%2Fconvert%2Fparquet/wikitext-2-v1/train/0000.parquet)

**Command:** `python convert.py -i convert/NeuralHermes-2.5-Mistral-7B -c convert/0000.parquet -o convert/temp2 -cf convert/nh-5bpw -b 5.0 -hb 6`

Layer measurements are provided in `measurement.json`` for further quantisation.

---

<center><img src="https://i.imgur.com/qIhaFNM.png"></center>

# NeuralHermes 2.5 - Mistral 7B

NeuralHermes is an [OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B) model that has been further fine-tuned with Direct Preference Optimization (DPO) using the [mlabonne/chatml_dpo_pairs](https://huggingface.co/datasets/mlabonne/chatml_dpo_pairs) dataset.

It is directly inspired by the RLHF process described by [neural-chat-7b-v3-1](https://huggingface.co/Intel/neural-chat-7b-v3-1)'s authors to improve performance. I used the same dataset and reformatted it to apply the ChatML template. I haven't performed a comprehensive evaluation of the model, but it works great, nothing broken apparently! :)

The code to train this model is available on [Google Colab](https://colab.research.google.com/drive/15iFBr1xWgztXvhrj5I9fBv20c7CFOPBE?usp=sharing) and [GitHub](https://github.com/mlabonne/llm-course/tree/main). It required an A100 GPU for about an hour.

GGUF versions of this model are available here: [mlabonne/NeuralHermes-2.5-Mistral-7B-GGUF](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B-GGUF).

## Usage

You can run this model using [LM Studio](https://lmstudio.ai/) or any other frontend.

You can also run this model using the following code:

```python
import transformers
from transformers import AutoTokenizer

# Format prompt
message = [
    {"role": "system", "content": "You are a helpful assistant chatbot."},
    {"role": "user", "content": "What is a Large Language Model?"}
]
tokenizer = AutoTokenizer.from_pretrained(new_model)
prompt = tokenizer.apply_chat_template(message, add_generation_prompt=True, tokenize=False)

# Create pipeline
pipeline = transformers.pipeline(
    "text-generation",
    model=new_model,
    tokenizer=tokenizer
)

# Generate text
sequences = pipeline(
    prompt,
    do_sample=True,
    temperature=0.7,
    top_p=0.9,
    num_return_sequences=1,
    max_length=200,
)
print(sequences[0]['generated_text'])
```


## Training hyperparameters

**LoRA**:
* r=16,
* lora_alpha=16,
* lora_dropout=0.05,
* bias="none",
* task_type="CAUSAL_LM",
* target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']

**Training arguments**:
* per_device_train_batch_size=4,
* gradient_accumulation_steps=4,
* gradient_checkpointing=True,
* learning_rate=5e-5,
* lr_scheduler_type="cosine",
* max_steps=200,
* optim="paged_adamw_32bit",
* warmup_steps=100,

**DPOTrainer**:
* beta=0.1,
* max_prompt_length=1024,
* max_length=1536,